The Stanford Center for Systems Biology of Cancer (CCSB) aims to discover molecular mechanisms underiying cancer progression by studying cancer as a complex biological system that is driven, in part, by impaired differentiation. Increasing evidence indicates that many cancers, like normal tissue, are composed of a hierarchy of cells at different stages of differentiation, and that the disease is maintained hy a self-renewing subpopulation. Our overarching goal is to provide a better understanding of the self-renewing properties of cancer that will enable us to identify molecular therapeutic targets and strategies to eradicate this disease, or to maintain it in a nonlethal state. Our biological projects are integrated with novel computational techniques, designed to dissect processes and causal factors underlying impaired differentiation as a driver of cancer progression in several hematologic malignancies. This approach will enable us to ascertain differences between these malignancies, and commonalities which may generalize to other cancers. In order to identify mechanistic underpinnings of cancer progression, a network-based and multiscale viewpoint is mandatory. Increasingly, diseases such as cancer are recognized as resulting from disruption in the coordinated performance of a complex biological system. This systems biology viewpoint necessitates the incorporation of high throughput, high dimensional data, and development of computational methods specifically geared to its analysis. There are three essential and interiocking requirements for a comprehensive systems analysis of cancer. First, powerful methods are required to infer molecular regulatory networks that drive phenotypic processes such as differentiation. Second, computational approaches are needed that can identify and isolate underlying patterns of progression in cancer, which can then be related to underlying regulatory networks. Third, executable models are desirable so that it is possible to pose hypothetical "what if' questions to predict how, for example, a targeted intervention might affect the subsequent course of disease. The approaches we will develop as a CCSB target these three specific computational aims. They are tailored to address the biological systems we are studying in our overall CCSB goal to understand the role of differentiation and self-renewal cancer. However, they will have much wider applicability. Thus, although here we apply them to particular biological systems, experimental testing of model predictions will validate not only the biological conclusions, but also the methodologies themselves. Furthermore, experimental validation will play a crucial role in iteratively refining and improving our computational models. Hematologic malignancies provide a unique opportunity to study the role of self-renewal and differentiation in cancer. Cells ofthe immune system develop from hematopoietic stem cells (HSCs) by a hierarchical process of differentiation to more specialized cell types, that has been well defined and studied. Self-renewing HSCs give rise initially to multipotent progenitors (MPPs) that have the potential to differentiate into multiple cell types, but lack self-renewal capacity. MPPs in tum give rise to oligopotent Common Myeloid Progenitor (CMP) and Common Lymphoid Progenitor (CLP), generating the major myeloid and lymphoid lineages that comprise the immune system. Subsequent differentiation produces progressively more specialized cell types that lack self-renewal ability, ultimately resulting in the major effector cells such as T-cells, B-cells, macrophages, and granulocytes. We will dissect the processes leading to deregulated differentiation, and acquisition of aberrant self-renewal ability in both myeloid and lymphoid lineages. For this purpose we will investigate three complementary systems: human Acute Myeloid Leukemia (AML), human Follicular Lymphoma (FL), and human and mouse T-cell Acute Lymphoblastic Lymphoma (T-ALL). Our computational methods produce network-level representations of molecular and cellular interactions that integrate diverse data types across multiple scales (molecular, cellular phenotypes, tumor phenotype, clinical outcomes) and filter the results through the viewpoint of differentiation and self-renewal pathways. By combining experimental and computational methods, we aim to predict and validate the critical aberrant molecular events that establish and maintain the self-renewal capacity of cancer, and how they relate to differentiation in normal cellular hierarchies. Our approaches are based on machine learning, executable models, multiscale modeling, and methods from the mathematics of geometry and topology. There will be a close interaction with experimental projects, in an iterative process where biological validation of computational predictions provides the basis for improved computational models. For this reason, computational methods development will occur under one project that interacts closely with all the experimental groups in our CCSB. The Stanford CCSB represents an evolution from our current status as a U56 ICBP Planning Center. In our cross-species systems biology analysis FL transformation and transgenic mouse models, the role of differentiation (and particulariy the aberrant activation of self-renewal programs) emerged as a key unifying theme in cancer progression. This proposal builds on our findings. We will extend our integrated systems studies into the role of differentiation and self-renewal in cancer, and how normal regulatory networks governing these processes become deregulated in cancer.